my prefect server setup prefect-metrics.waow.tech
python orchestration
0
fork

Configure Feed

Select the types of activity you want to include in your feed.

fix scoring: additive blend so new items aren't zeroed out

the multiplicative formula (recency * engagement * ...) gave score=0
to any item with no comments/reactions, burying 60% of open issues
including well-filed bugs. switch to additive blend where recency is
the primary signal and engagement boosts rather than gates.

140 → 14 zero-score items on current data.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

+15 -4
+7 -1
analytics/models/enrichment/int_github_issues_scored.sql
··· 17 17 LEFT JOIN {{ ref('known_contributors') }} kc ON i."user" = kc.login 18 18 ) 19 19 SELECT *, 20 - ROUND(recency_score * engagement_score * label_multiplier * contributor_weight, 4) AS importance_score 20 + -- additive blend: recency is primary signal, engagement boosts 21 + ROUND( 22 + (0.7 * recency_score + 0.3 * engagement_score) 23 + * label_multiplier 24 + * contributor_weight, 25 + 4 26 + ) AS importance_score 21 27 FROM scored 22 28 WHERE state = 'open'
+8 -3
analytics/models/enrichment/int_tangled_items_scored.sql
··· 3 3 t.*, 4 4 -- recency: decay over 30 days, max 1.0 5 5 GREATEST(0.0, 1.0 - DATEDIFF('day', t.created_at::DATE, CURRENT_DATE) / 30.0) AS recency_score, 6 - -- no engagement data on PDS — flat 0.5 7 - 0.5 AS engagement_score, 6 + -- no engagement data on PDS — flat 0.0 7 + 0.0 AS engagement_score, 8 8 -- contributor weight (from seed) 9 9 COALESCE(kc.weight, 1.0) AS contributor_weight 10 10 FROM {{ ref('stg_tangled_items') }} t 11 11 LEFT JOIN {{ ref('known_contributors') }} kc ON t.author_handle = kc.login 12 12 ) 13 13 SELECT *, 14 - ROUND(recency_score * engagement_score * contributor_weight, 4) AS importance_score 14 + -- same additive blend as github; engagement=0 means recency-only 15 + ROUND( 16 + (0.7 * recency_score + 0.3 * engagement_score) 17 + * contributor_weight, 18 + 4 19 + ) AS importance_score 15 20 FROM scored